I’m trying to connect graylog to the AWS kafka MKS and getting this error:
Input xxxxxxxxxxxxxxx has failed to start on node xxxxxxxxxxxxxxxxxxxxxx for this reason: »aws_kafka_001_ip-xxxxxxxxxxxxxxxx can’t rebalance after 4 retries.«. This means that you are unable to receive any messages from this input. This is mostly an indication for a misconfiguration or an error.
Judging from your error message something seams not to be configured correctly.
Is it possible to show any of your configuration?
Maybe the version of Graylog server or any other environment variables that you have setup?
rebalance.backoff.ms defines the time for which Kafka will wait before rebalancing & zookeeper.session.timeout.ms parameter specifies the time for which Kafka will wait to connect to zookeeper. So, try adjusting these two parameters.
Out of curiosity have you tried different port for the kafka input and/or possible with a different Input?
Do you see anything in your journal that may pertain to tis issue?
I have another Kafka input running on this graylog but that Kafka is not the MKS (aws kafka) and it is working as expected. So pretty much the only difference between the 2 inputs is that one is getting data from a Kafka cluster that is maintained by me and the input failing is the one that I’m trying to configure to get data from MKS. Not sure if there is a compatibility issue or something I don’t know about.
Is it possible to show a logical diagram on what your trying top accomplish? What I get of all this is that you created a two Inputs that are the same. One INPUT is running fine and the other INPUT one is not.
By chance have you tried to use a different type of Input. To be honest I’m not sure what Input you are using.